From 6d76e7d343b09d1a05823c2638fac1ebfc84098a Mon Sep 17 00:00:00 2001 From: bjf-frz Date: Wed, 22 Apr 2026 11:09:27 +0800 Subject: [PATCH] [cherry-pick][release/v.0.18.0.post1]remove duplicate video preprocess in Wan2.2 pipeline (#3005) Signed-off-by: bjf-frz --- vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2.py | 7 ------- vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_i2v.py | 7 ------- vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_ti2v.py | 7 ------- 3 files changed, 21 deletions(-) diff --git a/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2.py b/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2.py index 4b8d01dfa..22e0ed6b7 100644 --- a/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2.py +++ b/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2.py @@ -148,9 +148,6 @@ def get_wan22_pre_process_func( ): """Pre-process function for Wan2.2: optionally load and resize input image for I2V mode.""" import numpy as np - from diffusers.video_processor import VideoProcessor - - video_processor = VideoProcessor(vae_scale_factor=8) def pre_process_func(request: OmniDiffusionRequest) -> OmniDiffusionRequest: for i, prompt in enumerate(request.prompts): @@ -194,10 +191,6 @@ def get_wan22_pre_process_func( ) prompt["multi_modal_data"]["image"] = image # type: ignore # key existence already checked above - # Preprocess for VAE - prompt["additional_information"]["preprocessed_image"] = video_processor.preprocess( - image, height=request.sampling_params.height, width=request.sampling_params.width - ) request.prompts[i] = prompt return request diff --git a/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_i2v.py b/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_i2v.py index f9d063d6d..8d071f176 100644 --- a/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_i2v.py +++ b/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_i2v.py @@ -99,9 +99,6 @@ def get_wan22_i2v_pre_process_func( od_config: OmniDiffusionConfig, ): """Pre-process function for I2V: load and resize input image.""" - from diffusers.video_processor import VideoProcessor - - video_processor = VideoProcessor(vae_scale_factor=8) def pre_process_func(request: OmniDiffusionRequest) -> OmniDiffusionRequest: for i, prompt in enumerate(request.prompts): @@ -147,10 +144,6 @@ def get_wan22_i2v_pre_process_func( ) prompt["multi_modal_data"]["image"] = image # type: ignore # key existence already checked above - # Preprocess for VAE - prompt["additional_information"]["preprocessed_image"] = video_processor.preprocess( - image, height=request.sampling_params.height, width=request.sampling_params.width - ) request.prompts[i] = prompt return request diff --git a/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_ti2v.py b/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_ti2v.py index 4953a809d..884949b0b 100644 --- a/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_ti2v.py +++ b/vllm_omni/diffusion/models/wan2_2/pipeline_wan2_2_ti2v.py @@ -85,9 +85,6 @@ def get_wan22_ti2v_pre_process_func( od_config: OmniDiffusionConfig, ): """Pre-process function for TI2V: optionally load and resize input image.""" - from diffusers.video_processor import VideoProcessor - - video_processor = VideoProcessor(vae_scale_factor=8) def pre_process_func(request: OmniDiffusionRequest) -> OmniDiffusionRequest: for i, prompt in enumerate(request.prompts): @@ -133,10 +130,6 @@ def get_wan22_ti2v_pre_process_func( ) prompt["multi_modal_data"]["image"] = image # type: ignore # key existence already checked above - # Preprocess for VAE - prompt["additional_information"]["preprocessed_image"] = video_processor.preprocess( - image, height=request.sampling_params.height, width=request.sampling_params.width - ) request.prompts[i] = prompt return request